Richland & Kabini rumours

[NTMBK]

Sorry Krumme, I know it sounds easier to blame Intel for all AMD's woes but in this case wouldn't the most plausible explanation for not finding Brazos in models smaller than 15.6 be AMD's own restriction?

Llano and Trinity weren't exactly stars in sales performance. By putting Brazos in 13/14 notebooks it would cannibalize not only Intel sales, but also AMD sales in only of the few segments where it could get descent margins with Piledriver chips.

Bobcat was made for small laptops and netbooks in the first place. I sincerely doubt that AMD would put restrictions on its intended use.

 

[Vesku]

I don't think AMD has the luxury of putting restrictions on what their chips go into.

 

[mrmt]

Bobcat was made for small laptops and netbooks in the first place. I sincerely doubt that AMD would put restrictions on its intended use.

So you are saying that AMD should let Brazos cannibalize Llano/Trinity sales, worsening the WSA situation?

 

[mrmt]

I don't think AMD has the luxury of putting restrictions on what their chips go into.

It's not really a luxury, it's a necessity. Brazos is a variable cost for them. They only pay what they ask for and if they don't screw up on sales forecasting, they will pay for what they ask from TSMC.

Trinity is a little different. As they are below the WSA threshold, they have to pay for 265 million per quarter (COGS) of Trinity chips, whether they sell them or not. It's already sunk cost for them, so they have to sell or swallow the loss (pay the charge as they did last quarter) or inventory ballooning (as happened in Q2).

When faced with this situation, it would make sense to lose some Brazos sales to sell some Llano/Trinity, even if not at 1:1 ratio.

 

[Vesku]

I don't see why they would pay GloFo so they could make Kabini and Temash at TSMC just to then slow sell them. Better to work separately on pushing Trinity where Kabini and Temash just won't cut it.

 

[mrmt]

I don't see why they would pay GloFo so they could make Kabini and Temash at TSMC just to then slow sell them. Better to work separately on pushing Trinity where Kabini and Temash just won't cut it.

Let me try again:

- If AMD does not sell Brazos, they just don't order Brazos from TSMC and that's all.

- If AMD does not sell Trinity, they still have to pay GLF regardless of having ordered or not the chip.

So in a scenario where AMD had to chose between having to chose between selling Brazos and Trinity, it makes sense to not sell Brazos because they don't have to pay TSMC the chip. They are paying for one chip and selling one chip.

If they chose to sell Brazos but not Trinity, they still have to pay GLF for Trinity, because they have a take-or-pay clause in the WSA, so they are selling one chip but paying for two chips.

 

[frozentundra123456]

As we can see from the bobcat apearance on laptops Intel will try the folowing by kickbacks, offers, scaring/motivating ...:

Dont let kabini into anything smaller than an 15.6
Dont let kabini use its faster models

For AMD the response this time should be very simple. Only two variants of kabini, no scaled down versions:

Quad core 15w
Quad core 25w

Simple as that. Take it or leave it. For the few kabinies that down meet the Ghz, burn them. Better than burn your brand, even though there is hardly any left. Notebookcheck will fx. take whatever low performance variant comes out and compare it to an high-perf ulv IB straigth away. And we will see tons of stupid comparison also here on anandtech - can you say single threaded performance ? - compared to the core line, and low end discrete Nivida/AMD cards - even though its completely different markets. Thats the name of the game when you are the small player, and therefore, you have to be more selective what you put on the market. AMD havnt been that.

Use the same tactics for Temash. No discount, only high performance/efficiency parts. For the embedded market - tv streamer and so on - its a different story. But they need to man up here, otherwise they will automaticly be positioned by Intel marketing/sale to the lower side. The strategy here, i think is in line with Pablos more push approach.

So now notebookcheck and anand tech are also against amd? OK.............

 

[Idontcare]

So now notebookcheck and anand tech are also against amd? OK.............

Its just confirmation bias of another color. Don't like the answer then seek to discredit the source.

I remember when the shoe was on the other foot, in the netburst era, and the Intel fanboys decried pro-AMD bias of Anand in his CPU reviews then. Same psychological issue then as now.

People want to see their preferred position supported, they seek links to support it (confirmation bias) and deride the credibility of anything that would otherwise undermine their preference.

Nothing to be done about it, it'll be the same in 5yrs too.

 

[MeldarthX]

IDC - I remember that and several other times - but some of the testing you do have scratch your head at LOL......

I've been a lurker forever; I remember the forums before they switched...

But I think Intel is nervous about Kabini and Temash; but if Kabini had come out when it was supposed to; I think it actually terrified Intel back then; now just nervous.....

While the anti trust in the US is closed but not on Intel; in EU its a completely different matter; the investigation is on going and they have to tread carefully. I think here in the EU and UK we tend to see more models of AMD systems; specially HP; Lenova; Acer than the states do. That is because Intel.....

I do think RR is making head way; specially after AMD scoring all the next gen Consoles; there working much more with the developers with Games....we're starting to see the newer game engines actually use the cores.

I think once we start getting benches from Kabini and Temash; if they are kicked the crap out of Intel equivalent but end up in almost none or no products then we got to ask ourselves why. Is it AMD or is it Intel again playing dirty?

I've seen the tactics first hand as I was build engineer and Talon production manager at Falcon NW from 1999-2004. I know the deals that Intel struck to kill us from using the fastest Athlons over the P4 *which at the time blew* Then also tried to keep us from using X2s......

 

[Vesku]

Let me try again:

- If AMD does not sell Brazos, they just don't order Brazos from TSMC and that's all.

- If AMD does not sell Trinity, they still have to pay GLF regardless of having ordered or not the chip.

So in a scenario where AMD had to chose between having to chose between selling Brazos and Trinity, it makes sense to not sell Brazos because they don't have to pay TSMC the chip. They are paying for one chip and selling one chip.

If they chose to sell Brazos but not Trinity, they still have to pay GLF for Trinity, because they have a take-or-pay clause in the WSA, so they are selling one chip but paying for two chips.

They don't pay by the chip as it comes off the line to TSMC, they still have to order wafers in advance of sales. They need to drive up demand of any and all products. I'm sure they'd rather kabini and temash are smash hits and pay off GloFo then to be stuck with more inventory to write off. Otherwise they would have just taken the punches and rolled out the crappier 32nm GloFo versions (wichita and something else?).

 

[Olikan]

But I think Intel is nervous about Kabini and Temash; but if Kabini had come out when it was supposed to; I think it actually terrified Intel back then; now just nervous.....

some weeks ago i read somewhere that some Intel guys were a bit terrified with kaveri...but then AMD delayed it, now they just laught

 

[mrmt]

They don't pay by the chip as it comes off the line to TSMC, they still have to order wafers in advance of sales. They need to drive up demand of any and all products. I'm sure they'd rather kabini and temash are smash hits and pay off GloFo then to be stuck with more inventory to write off. Otherwise they would have just taken the punches and rolled out the crappier 32nm GloFo versions (wichita and something else?).

Per chip, per wafer, it doesn't matter. AMD has control over their costs with TSMC, the same can't be said of GLF. Until they hit the WSA threshold, selling
$1 worth of Trinity and selling $265 million worth of Trinity chips costs AMD $265 million, so they need to push Trinity sales the closest they can to the $265 million threshold.

What you are saying is what you are "sure", but I'm yet to see the math beyond your gut feeling. Care to explain.

 

[krumme]

Sorry Krumme, I know it sounds easier to blame Intel for all AMD's woes

Dont use that strawman on me. Every executive of a large corp. know what it means to be the big fish in the pond and what a pain it is to be nr. 2. I am used to be the with the big fish and i know what advantage it is. If you think my commets are harsh, they are now where near the brutal reality. Go read the original pdf for NY lawsuit against Intel. http://semiaccurate.com/forums/showpost.php?p=11329&postcount=3

Here are some point:

Guess When Dell got this?

“Stop writing checks immediately and
put them back on list prices asap.”
..
Otellini reported back on a telephone conversation with Dell’s
CEO Kevin Rollins:
I had my call with Kevin yesterday. It went well. He did NOT ask for money …
he called to …

...
In this email, Otellini wrote:
- [Michael Dell] opened by saying “I am tired of losing business” … he
repeated it 3-4 times. I said nothing and waited.
- He has been traveling around the USA. He feels they are losing all the high
margin business to AMD-based sku’s …
- He is ‘tired of being behind for 4 years (when I protested that it was 2, he
said, no the last 2 years, this year, and next year).
- As a result, “Dell is no longer seen as a thought leader”

...
On November 10, 2005, Michael Dell followed up with an email to Otellini: “We
have lost the performance leadership and it’s seriously impacting our business in several areas.”
Otellini’s reply: “There is nothing new here. Our product roadmap is what it is. It is improving
rapidly daily. It will deliver increasingly leadership products … Additionally, we are
transferring over $1B per year to Dell for meet comp efforts. This was judged by your team to be more than sufficient to compensate for the competitive issues.”

.....

On February 16, 2006, Intel took note of a service report in which Dell’s CEO
Kevin Rollins had said that Dell had “made no plans to begin using” AMD chips. “Finally
something positive” commented one Intel executive. Otellini commented: “The best friend
money can buy.” (Emphasis added).
"

Llano and Trinity weren't exactly stars in sales performance. By putting Brazos in 13/14 notebooks it would cannibalize not only Intel sales, but also AMD sales in only of the few segments where it could get descent margins with Piledriver chips.

Remember that until they are manufacturing Jaguar or successors in GLF, they must sell at least 265 million (COGS) in Piledriver chips per quarter.

I understand your argument, but trinity was hardly in any 11.6/13 at all, so the argument is not valid.

There is a reason you pay for PR, marketing, technical marketing and huge sales force. It gives results. When do people understand that in these forums. They think its always pure overhead. In their world its the fully informed buyer making rational choices, like there is only logictics between production and usage.

 

[MeldarthX]

some weeks ago i read somewhere that some Intel guys were a bit terrified with kaveri...but then AMD delayed it, now they just laught

No everything I've seen and read; they aren't terrified of it anymore but still nervous of it......

specially since Intel's drivers; and PQ are udder and complete crap

AMD got a great op to become a house hold name with the new consoles - their logo on the box....but we shall see....

 

[sequoia464]

- If AMD does not sell Trinity, they still have to pay GLF regardless of having ordered or not the chip.

How long is this in force?

 

[mrmt]

Dont use that strawman on me. Every executive of a large corp. know what it means to be the big fish in the pond and what a pain it is to be nr. 2.

Not trying to make you that, sorry.

I understand your argument, but trinity was hardly in any 11.6/13 at all, so the argument is not valid.

http://forum.notebookreview.com/har...76413-ultimate-amd-trinity-notebook-list.html

Plenty of 13/14 inch Trinity here. I don't know if AMD would trade those sales for Brazos.

 

[Vesku]

The math:

-$100s of millions, Cancel Wichita and Krishna, delay whole product segment by a year
-$320 million, To avoid bringing the retrenched replacements for Wichita and Krishna over to GlobalFoundries

-$420+ million all so that Kabini and Temash could arrive. They need Kabini and Temash to be a success, they want kabini and temash to be a success, I just can't imagine AMD doing anything that might cool customers from those chips.

If they just wanted to not get dinged by their crazy WSA (I mean excluding a one time negotiated exception they are paying full price even when yields and performance is lackluster) with their former foundry they would have just pushed out Wichita and Krishna as mediocre products. On GloFo 32nm if 28nm was unusable. But they made costly choices to not do that.

*Note: I'm not sure where "here" is but on Newegg there are 4 non-refurbished AMD A series notebooks in the 13-14 inch category. 3 of them are variations of a single Lenovo model type.

 

[mrmt]

-$320 million, To avoid bringing the retrenched replacements for Wichita and Krishna over to GlobalFoundries

The 320 million charge wasn't paid to leave Kabini at TSMC, as the decision to manufacture it there was taken at least in Q112, not Q412. This charge was paid because AMD could not sell enough Trinity/Bulldozer/Piledriver, it had *nothing* to do with Kabini or Brazos.

And btw, the exclusivity waiver was cancelled, all future CPU products from AMD will be manufactured at GLF including Kabini successor if not Kabini itself.

 

[krumme]

So now notebookcheck and anand tech are also against amd? OK.............

Where did i write that?............
They have a business. Anand places add handling at an exterior company, to secure no influence. He have very strict rules compared to man others. But every site like anand live of information, and therefore good connections. Thats why we come to Anandtech, and it gives us good insight.
But do you truly beliewe Intel or AMD does that because they like him or us? - they "use" him for the same reason he uses them. Its okey, its managed, but it always creates bias, thats whats pays for the show. Better know that.

 

[Vesku]

The 320 million charge wasn't paid to leave Kabini at TSMC, as the decision to manufacture it there was taken at least in Q112, not Q412. This charge was paid because AMD could not sell enough Trinity/Bulldozer/Piledriver, it had *nothing* to do with Kabini or Brazos.

And btw, the exclusivity waiver was cancelled, all future CPU products from AMD will be manufactured at GLF including Kabini successor if not Kabini itself.

They could have avoided that $320 million charge, they saw it coming. They could have resurrected Wichita and Krishna and met their GloFo obligations. They chose to pay out. AMD's plan is to bring it all back in to GloFo when 22nm is up and running which so far means the 2014 generation of chips. That GloFo 22nm won't be a drag on their product lines like 32nm ramp up and discontinued 28nm, I wish you luck with that AMD.

 

[Phynaz]

Go read the original pdf for NY lawsuit against Intel. http://semiaccurate.com/forums/showpost.php?p=11329&postcount=3

What was the verdict?

 

[Vesku]

Verdict was, rich entity hands struggling entity a wad of cash to reach a settlement.

Honestly given Intel's x86 market position those OEM emails were quite damning. However, even governments are hesitant to go after the very wealthy because the court cases can go on for ages.

I believe the direct payola has been mostly curtailed now and considering AMD accepted the settlement money, pointing fingers at this juncture is some sort of zen forum exercise.

Current AMD problems are mostly to do with production and failure to form their own strong relationships with customers. Granted Intel probably still plays a factor in the latter, as in what x86 maker would want to get on Intel's bad side? But with payola curtailed that's just plain business reality.

 

[mrmt]

They could have avoided that $320 million charge, they saw it coming. They could have resurrected Wichita and Krishna and met their GloFo obligations.

They saw? Really? So it was just AMD executive team making fun of themselves in a game to bring the market cap from over 6 billion to less than 2 just to save Kabini? Because they have been missing their sales projection for the entire 2012, and guess what, every time they missed it was for selling less than forecast, not more.

And AMD is planning to move to the subpar foundry in an even riskier process than the vanilla 28nm GLF is in trouble to make it work?

And how can you say thay they could resurrect Krishna and Wichita when only in Q113 28nm at GLF was ready for production? This admission came from GLF itself.

 

[krumme]

Not trying to make you that, sorry.



http://forum.notebookreview.com/har...76413-ultimate-amd-trinity-notebook-list.html

Plenty of 13/14 inch Trinity here. I don't know if AMD would trade those sales for Brazos.

They were, but they were not sold, as the price was - as someone else wrote - way to high.

Anyway thats history. Trinity is a bad product, especially for anything under 25W, and bobcat was way to weak at 9w and below. Jaguar is much better because of scalability, but have a far shorter window of oportunity.

AMD have a window here 2013, q1 2014 perhaps q2. If they can not use it, i have a hard time seeing how the 2B market should be justified. Then they are just deep under water, and will have trouble ajusting cost fast enough. They hardly have the money to be in this business, and it can easily go worse.

For some reason it have always been like that for AMD, constantly losing money, and somehow they are still here.

 

[Vesku]

They saw? Really? So it was just AMD executive team making fun of themselves in a game to bring the market cap from over 6 billion to less than 2 just to save Kabini? Because they have been missing their sales projection for the entire 2012, and guess what, every time they missed it was for selling less than forecast, not more.

And AMD is planning to move to the subpar foundry in an even riskier process than the vanilla 28nm GLF is in trouble to make it work?

And how can you say thay they could resurrect Krishna and Wichita when only in Q113 28nm at GLF was ready for production? This admission came from GLF itself.

Late 2011 - AMD sees that GloFo 28nm is as much if not more of a mess than 32nm.

Decision point: Rework Wichita and Krishna for 32nm GloFo or 28nm TSMC.

28nm TSMC, Known Consequences: Those wafers won't count towards WSA with GloFo

Now, do you expect me to believe that AMD management team thought the sales of Trinity would be enough over sales of Llano to cover leaving their Brazos segment on TSMC? Given that most PC related companies were projecting a slowdown in 2012? Keep in mind their special pricing deal for a poor 32nm rollout was going away as well.

Final note: If it wasn't for their agreement with GloFo do you think they'd be moving everything to GloFo 22nm? Given how 32nm and 28nm has done?

PR is a funny beast, here is GloFo advertising how a tiny IP company is making their IP available on their 28nm. http://www.globalfoundries.com/newsroom/2013/20130205-Adapteva.aspx
Just so they can pretend there is some interest.